Model Context Protocol (MCP) adopted by AI assistants and developer tools for decentralized search

Introduction to Model Context Protocol (MCP)

The Model Context Protocol (MCP) is an open protocol designed to change how AI assistants retrieve information. Instead of relying on internal model memories, MCP allows assistants to call tools provided by a server to search, read resources, and fill text templates. This means information is fetched directly from the publishers, ensuring it is current and sourced.

Protocol Adoption and Governance

Anthropic launched MCP in November 2024. Within months, major AI assistants and developer tools adopted the protocol. In December 2025, MCP was transferred to the Agentic AI Foundation, which operates under the Linux Foundation. The protocol evolves through dated versions, and a public registry lists available servers.

Decentralized Search Mechanism

Unlike traditional search engines that index copies of web content, MCP operates differently. Publishers maintain their data on their own servers and expose it through these servers. AI assistants query multiple servers on demand, retrieving answers that include the source, license, author, and last update date. This approach decentralizes search, allowing various entities like statistics offices or specialized websites to directly provide up-to-date information.

Benefits for Users

Users benefit from MCP through up-to-date answers, as assistants read data directly from servers at the time of the query. Answers come with verifiable sources, linking back to original pages. Clear rights are provided, as server-stated licenses travel with the data. This system also aims to reduce AI 'invention' by allowing servers to indicate when information is unavailable, potentially leading to fewer fabricated responses.